Thursday, 2019-08-22

*** zaitcev has quit IRC00:03
*** pvradu has joined #openstack-swift00:12
*** zaitcev has joined #openstack-swift00:17
*** ChanServ sets mode: +v zaitcev00:17
*** takamatsu has quit IRC00:17
*** gyee has quit IRC00:32
*** zaitcev has quit IRC00:55
*** hogepodge has quit IRC00:55
*** csmart has quit IRC00:55
*** ChanServ has quit IRC00:55
*** tkajinam has quit IRC00:55
*** threestrands has quit IRC00:55
*** clayg has quit IRC00:55
*** nottrobin has quit IRC00:55
*** patchbot has quit IRC00:55
*** rcernin has quit IRC00:55
*** DHE has quit IRC00:55
*** baojg has quit IRC00:55
*** mrjk has quit IRC00:56
*** cwright has quit IRC00:56
*** joeljwright has quit IRC00:56
*** stamps has quit IRC00:56
*** evrardjp has quit IRC00:56
*** lifeless has quit IRC00:56
*** sorrison has quit IRC00:56
*** camelCaser has quit IRC00:56
*** kukacz has quit IRC00:56
*** cschwede has quit IRC00:56
*** early has quit IRC00:56
*** notmyname has quit IRC00:56
*** kota_ has quit IRC00:56
*** timss has quit IRC00:56
*** hoonetorg has quit IRC00:56
*** openstackgerrit has quit IRC00:56
*** dasp has quit IRC00:56
*** tonyb has quit IRC00:56
*** MooingLemur has quit IRC00:56
*** irclogbot_3 has quit IRC00:56
*** dosaboy has quit IRC00:56
*** ndk_ has quit IRC00:56
*** paladox has quit IRC00:56
*** Anticimex has quit IRC00:56
*** hugokuo has quit IRC00:56
*** fungi has quit IRC00:56
*** seongsoocho has quit IRC00:56
*** fyx has quit IRC00:56
*** godog has quit IRC00:56
*** mugsie has quit IRC00:56
*** gmann has quit IRC00:56
*** donnyd has quit IRC00:56
*** jungleboyj has quit IRC00:56
*** rledisez has quit IRC00:56
*** alecuyer has quit IRC00:56
*** pvradu has quit IRC00:56
*** jistr has quit IRC00:56
*** rickflare has quit IRC00:56
*** mahatic has quit IRC00:56
*** corvus has quit IRC00:56
*** kazsh has quit IRC00:56
*** mattoliverau has quit IRC00:56
*** coreycb has quit IRC00:56
*** d34dh0r53 has quit IRC00:56
*** abelur has quit IRC00:56
*** StevenK has quit IRC00:56
*** beisner has quit IRC00:56
*** f0o has quit IRC00:56
*** mgagne has quit IRC00:56
*** onovy has quit IRC00:56
*** Jeffrey4l has quit IRC00:56
*** redrobot has quit IRC00:56
*** spotz has quit IRC00:56
*** szaher has quit IRC00:56
*** dabukalam has quit IRC00:56
*** timur has quit IRC00:56
*** timur has joined #openstack-swift01:14
*** szaher has joined #openstack-swift01:14
*** spotz has joined #openstack-swift01:14
*** redrobot has joined #openstack-swift01:14
*** Jeffrey4l has joined #openstack-swift01:14
*** dabukalam has joined #openstack-swift01:14
*** sorrison has joined #openstack-swift01:14
*** lifeless has joined #openstack-swift01:14
*** stamps has joined #openstack-swift01:14
*** evrardjp has joined #openstack-swift01:14
*** joeljwright has joined #openstack-swift01:14
*** cwright has joined #openstack-swift01:14
*** mrjk has joined #openstack-swift01:14
*** baojg has joined #openstack-swift01:14
*** DHE has joined #openstack-swift01:14
*** rcernin has joined #openstack-swift01:14
*** MooingLemur has joined #openstack-swift01:14
*** tonyb has joined #openstack-swift01:14
*** dasp has joined #openstack-swift01:14
*** openstackgerrit has joined #openstack-swift01:14
*** hoonetorg has joined #openstack-swift01:14
*** hogepodge has joined #openstack-swift01:14
*** zaitcev has joined #openstack-swift01:14
*** onovy has joined #openstack-swift01:14
*** mgagne has joined #openstack-swift01:14
*** timss has joined #openstack-swift01:14
*** kota_ has joined #openstack-swift01:14
*** notmyname has joined #openstack-swift01:14
*** card.freenode.net sets mode: +vvvv joeljwright zaitcev kota_ notmyname01:14
*** early has joined #openstack-swift01:14
*** cschwede has joined #openstack-swift01:14
*** kukacz has joined #openstack-swift01:14
*** camelCaser has joined #openstack-swift01:14
*** patchbot has joined #openstack-swift01:14
*** clayg has joined #openstack-swift01:14
*** nottrobin has joined #openstack-swift01:14
*** threestrands has joined #openstack-swift01:14
*** tkajinam has joined #openstack-swift01:14
*** f0o has joined #openstack-swift01:14
*** beisner has joined #openstack-swift01:14
*** StevenK has joined #openstack-swift01:14
*** abelur has joined #openstack-swift01:14
*** d34dh0r53 has joined #openstack-swift01:14
*** coreycb has joined #openstack-swift01:14
*** card.freenode.net sets mode: +vv cschwede clayg01:14
*** ChanServ has joined #openstack-swift01:14
*** card.freenode.net sets mode: +o ChanServ01:14
*** ndk_ has joined #openstack-swift01:14
*** paladox has joined #openstack-swift01:14
*** hugokuo has joined #openstack-swift01:14
*** jistr has joined #openstack-swift01:15
*** rickflare has joined #openstack-swift01:15
*** corvus has joined #openstack-swift01:15
*** mahatic has joined #openstack-swift01:15
*** kazsh has joined #openstack-swift01:15
*** mattoliverau has joined #openstack-swift01:15
*** card.freenode.net sets mode: +vv mahatic mattoliverau01:15
*** sorrison has quit IRC01:17
*** mugsie has joined #openstack-swift01:21
*** Anticimex has joined #openstack-swift01:21
*** persia has joined #openstack-swift01:22
*** sorrison has joined #openstack-swift01:23
*** gmann has joined #openstack-swift01:23
*** irclogbot_2 has joined #openstack-swift01:25
*** gmann has quit IRC01:27
*** gmann has joined #openstack-swift01:28
*** fungi has joined #openstack-swift01:37
*** ianychoi has quit IRC02:19
*** ianychoi has joined #openstack-swift02:19
*** baojg has quit IRC02:38
*** baojg has joined #openstack-swift02:44
*** baojg has quit IRC02:51
*** baojg has joined #openstack-swift02:53
*** BjoernT has joined #openstack-swift02:55
*** gkadam has joined #openstack-swift03:37
*** gkadam is now known as gkadam-afk03:38
*** gkadam has joined #openstack-swift04:17
*** gkadam-afk has quit IRC04:19
*** pvradu has joined #openstack-swift04:22
*** pvradu has quit IRC04:27
*** zaitcev has quit IRC04:46
*** csmart has joined #openstack-swift04:48
*** BjoernT_ has joined #openstack-swift04:49
*** BjoernT has quit IRC04:50
*** BjoernT_ has quit IRC05:15
openstackgerritMerged openstack/swift feature/losf: Merge branch 'master' into feature/losf  https://review.opendev.org/67784805:17
*** takamatsu has joined #openstack-swift05:47
*** ianychoi has quit IRC06:15
*** ianychoi has joined #openstack-swift06:19
*** pvradu has joined #openstack-swift06:24
*** ianychoi has quit IRC06:28
*** ianychoi has joined #openstack-swift06:29
*** ianychoi has quit IRC06:52
*** ianychoi has joined #openstack-swift06:54
*** takamatsu has quit IRC06:57
*** pvradu has quit IRC06:57
*** e0ne has joined #openstack-swift07:37
*** ianychoi has quit IRC07:52
*** ianychoi has joined #openstack-swift07:54
*** threestrands has quit IRC08:01
openstackgerritMerged openstack/swift master: py3: Finish porting func tests  https://review.opendev.org/67522708:13
*** takamatsu has joined #openstack-swift08:21
*** godog has joined #openstack-swift08:21
*** tkajinam has quit IRC08:23
*** tdasilva has joined #openstack-swift08:38
*** ChanServ sets mode: +v tdasilva08:38
*** pvradu has joined #openstack-swift08:42
*** tdasilva_ has joined #openstack-swift08:51
*** ChanServ sets mode: +v tdasilva_08:51
*** tdasilva has quit IRC08:52
*** tdasilva_ is now known as tdasilva08:52
*** rcernin has quit IRC08:54
*** tdasilva_ has joined #openstack-swift08:55
*** ChanServ sets mode: +v tdasilva_08:55
*** tdasilva_ has quit IRC08:55
*** tdasilva has quit IRC08:58
*** tdasilva has joined #openstack-swift09:03
*** ChanServ sets mode: +v tdasilva09:03
*** takamatsu has quit IRC09:08
*** e0ne has quit IRC09:42
*** e0ne has joined #openstack-swift09:56
*** takamatsu has joined #openstack-swift10:55
*** tesseract has joined #openstack-swift11:07
*** takamatsu has quit IRC11:25
*** takamatsu has joined #openstack-swift11:45
*** alecuyer has joined #openstack-swift12:00
*** rledisez has joined #openstack-swift12:01
*** ChanServ sets mode: +v rledisez12:01
*** ianychoi has quit IRC12:07
*** ianychoi has joined #openstack-swift12:09
openstackgerritThiago da Silva proposed openstack/swift master: symlink-backed versioned_writes  https://review.opendev.org/63385712:11
*** ianychoi has quit IRC12:27
*** ianychoi has joined #openstack-swift12:28
*** gkadam has quit IRC12:32
*** donnyd has joined #openstack-swift12:32
openstackgerritThiago da Silva proposed openstack/swift master: symlink-backed versioned_writes  https://review.opendev.org/63385712:59
*** BjoernT has joined #openstack-swift13:19
*** BjoernT_ has joined #openstack-swift13:33
*** BjoernT has quit IRC13:34
*** altlogbot_0 has joined #openstack-swift13:47
*** altlogbot_0 has quit IRC13:48
*** altlogbot_2 has joined #openstack-swift13:52
*** baojg has quit IRC14:09
*** spsurya has joined #openstack-swift14:10
*** baojg has joined #openstack-swift14:12
openstackgerritThiago da Silva proposed openstack/swift master: Add rebalance seed to build rings in docker image  https://review.opendev.org/67803414:41
*** takamatsu has quit IRC14:55
tdasilvatimburke: while testing p 678034, i noticed that i'm able to start and stop a py2 container as many times as I want and I can get data back, but if I try to "migrate" to a py3 container, i'm no longer able to. would you expect our ring rebalance to change even when using same seed value?15:12
patchbothttps://review.opendev.org/#/c/678034/ - swift - Add rebalance seed to build rings in docker image - 1 patch set15:12
timburketdasilva, yes :-(15:12
tdasilvascary :/15:13
timburkei remember trying incredibly hard in https://review.opendev.org/#/c/401397/ to get them to match, but ultimately i had to just give up so i could keep moving15:15
patchbotpatch 401397 - swift - py3: port common/ring/ and common/utils.py (MERGED) - 14 patch sets15:15
timburkei think patchset 7 had me trying to make py3 random behave like py2 random? but there were *still* issues15:18
timburketdasilva, so what if we think about it this way: why do we have to build the rings every time? why isn't that considered an output of the container?15:19
timburkelike, create it if it doesn't exist, but store it on the bind mount?15:20
tdasilvatimburke: yep! that was actually going to be a different approach I was going to try testing, but mostly for testing migration from py2 to py315:22
tdasilvafor the bind mounting of the data path, I just wanted to enable users to be able to save their data without thinking too much about swift configuration15:23
tdasilvaso still the same goal of just providing an API endpoint useful for application dev. testing, with the additional benefit of being able to keep data between container restarts15:24
tdasilvabut but but, for us swift devs, i was hoping it could be useful to testing py2-py3 migration for example, or just release upgrade testing15:24
tdasilvawdyt?15:27
timburkei think for app dev testing, the existing setup isn't so bad -- provided they just have the one "drive"15:27
timburkeit's all single-replica anyway, right?15:27
timburkefor the broader case... idk, i think we need to make whoever's running this start thinking more about configs15:28
timburkehmm. so if you really want to run containerized swift... do you have multiple drives per container? or do you just spin up a container per drive? my gut says the latter...15:31
tdasilvawell...right now as is, the container is a single saio, so we create 6 internal dirs and add 6 devices to a single replica ring which doesn't make much sense, i don't think...15:33
openstackgerritTim Burke proposed openstack/swift master: py3: Switch swift-dsvm-functional-py3 to run tests under py3  https://review.opendev.org/67704715:34
tdasilvahow if one was really serious about containerizing swift to run in prod, it might make sense to investgate running object servers with server per port enabled and run one object-server container per device...maybe??15:35
tdasilvas/how/Now15:35
*** gyee has joined #openstack-swift15:50
*** pvradu has quit IRC15:53
*** gkadam has joined #openstack-swift16:02
*** gkadam has quit IRC16:06
*** altlogbot_2 has quit IRC16:14
*** ianychoi has quit IRC16:15
*** altlogbot_1 has joined #openstack-swift16:15
*** ianychoi has joined #openstack-swift16:17
*** ianychoi has quit IRC16:19
*** ianychoi has joined #openstack-swift16:19
openstackgerritTim Burke proposed openstack/swift master: Autovivify X-Versions-Location container  https://review.opendev.org/26501516:23
*** e0ne has quit IRC16:54
*** psachin has joined #openstack-swift16:55
*** psachin is now known as psachin|away17:12
*** tesseract has quit IRC17:13
*** psachin|away has quit IRC17:32
*** psachin has joined #openstack-swift17:33
*** zaitcev has joined #openstack-swift17:38
*** ChanServ sets mode: +v zaitcev17:38
*** zaitcev has quit IRC17:43
*** psachin has quit IRC17:50
*** zaitcev has joined #openstack-swift17:54
*** ChanServ sets mode: +v zaitcev17:54
*** e0ne has joined #openstack-swift17:59
*** zaitcev has quit IRC18:01
*** e0ne has quit IRC18:11
*** e0ne has joined #openstack-swift18:11
*** irclogbot_2 has quit IRC18:14
*** e0ne has quit IRC18:14
*** zaitcev has joined #openstack-swift18:15
*** ChanServ sets mode: +v zaitcev18:15
*** irclogbot_2 has joined #openstack-swift18:16
*** altlogbot_1 has quit IRC18:29
*** e0ne has joined #openstack-swift18:52
*** spsurya has quit IRC18:52
*** zaitcev_ has joined #openstack-swift19:00
*** ChanServ sets mode: +v zaitcev_19:00
*** zaitcev has quit IRC19:03
tdasilvatimburke, clayg I’m still scratching my head over how to handle object expiration and object versions. The idea of expiring old versions seems counterintuitive19:09
tdasilvaThis is one area where I actually like the idea of the bucket policies where users intentionally would set a policy such as delete previous versions older than 30 days.19:11
timburketdasilva, i agree wholeheartedly -- i don't think any of the interactions there are intuitive or even necessarily well-understood (much less documented)19:49
timburkeso if anyone has opinions on how they *should* interact, maybe that'd be a better direction to start looking...19:50
timburkei *think* i know what i'd want it to look like with x-history-location -- record the delete marker and clean up the primary container. but i think trying to apply that to x-versions-location gets weird...19:52
timburke(there are some edge cases with history mode, of course -- like how to handle an overwrite between when the object expired and when the expirer got around to trying to clean it up. but especially with symlink-backed versioning, i think it's all resolvable)19:54
tdasilvatimburke: would blocking X-delete-at header on a versions container be a totally crazy idea?19:54
timburkewe still have to worry about it, at least somewhat. what if the container replicas are out-of-sync and don't agree about whether versioning is enabled or not? somebody somewhere is going to be able to get that expiry entry in the queue19:56
timburkethis is particularly likely if the client does a container PUT before starting a round of uploads -- which i'm pretty sure is also what led to https://bugs.launchpad.net/swift/+bug/183935519:57
openstackLaunchpad bug 1839355 in OpenStack Object Storage (swift) "container-sharder should keep cleaving when there are no rows" [Undecided,In progress] - Assigned to Matthew Oliver (matt-0)19:57
timburkei'm starting to understand why s3 treats a PUT to an already-existing container as a client error ;-)19:58
openstackgerritTim Burke proposed openstack/swift master: py3: Switch swift-dsvm-functional-py3 to run tests under py3  https://review.opendev.org/67704720:10
claygtdasilva: so, i guess in the old copy design, were we like forwarding x-delete-at headers across to old versions and now that's broken/different?20:10
clayglike if you uploaded an object x-delete-after: 30 days, then 3 days later you overwrite it - the backup version get's PUT with a new expirery queue entry pointing at it in the new location but with the delete-at carried forward (e.g. it'll expire after 27 days)20:11
claygcause we could probably make that behavior mostly work until an expiration makes a symlink broken20:12
*** e0ne has quit IRC20:56
claygdid rledisez make any progress on p 184078021:10
patchbotNo data found for patch 184078021:10
claygerrr.. lp bug #184078021:10
openstackLaunchpad bug 1840780 in OpenStack Object Storage (swift) "Eventlet exception "do not call blocking functions from the mainloop"" [Undecided,New] https://launchpad.net/bugs/184078021:10
*** BjoernT_ has quit IRC21:40
*** zaitcev_ has quit IRC22:09
*** zaitcev_ has joined #openstack-swift22:23
*** ChanServ sets mode: +v zaitcev_22:23
timburkegdi -- i hate it when i do a rebase, fix up the obvious conflicts, continue the rebase, then run tests under py3 to discover the *non-obvious* conflicts, get everything squared, git review... and then realize i forgot to commit and git review did a reset --hard :-(22:25
zaitcev_wait, git review can do that?!22:26
timburkehttps://opendev.org/opendev/git-review/src/branch/master/git_review/cmd.py#L88022:26
timburkei think it maybe has something to do with me configuring [rebase]autoStash = true ?22:28
openstackgerritTim Burke proposed openstack/swift master: Autovivify X-Versions-Location container  https://review.opendev.org/26501522:28
zaitcev_That's at the level of "del" with no arguments deleting the whole current directory.22:29
zaitcev_I think I read about it in "The Brief History of The Hackerdom".22:30
*** swifterdarrell has joined #openstack-swift22:36
*** ChanServ sets mode: +v swifterdarrell22:36
timburkehmm. do we have any probe tests that exercise versioning corner cases?22:38
*** swifterdarrell has quit IRC22:40
*** swifterdarrell has joined #openstack-swift22:40
*** ChanServ sets mode: +v swifterdarrell22:40
swifterdarrelltimburke: I remembered how IRC works!22:41
timburkeswifterdarrell \o/22:41
timburkenow i won't have to translate your findings for rledisez and alecuyer!22:41
*** zaitcev_ is now known as zaitcev22:50
*** rcernin has joined #openstack-swift22:55
*** rcernin is now known as rcernin|brb23:00
*** tkajinam has joined #openstack-swift23:03
*** rcernin|brb has quit IRC23:09
*** rcernin has joined #openstack-swift23:10
*** persia has left #openstack-swift23:55

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!